What is an Artifact?
A durable, shareable output created by an AI agent, such as a report, application, analysis, document, dashboard, or interactive interface.
Why is this important?
Artifacts separate useful work products from transient chat and create stable objects teams can review, share, version, and attach to context.
How it works
An agent generates structured content or code, stores it with metadata and permissions, renders it in a viewer, and preserves links to source records.
Technical example
A weekly pipeline review becomes an interactive report shared with a manager while retaining the permissions of the underlying CRM records.
Implementation notes
Record creator, sources, version, access scope, freshness, and whether the artifact can execute actions or only display information.
